In a computer network context, servers may be used provide network content. In some cases, the servers dynamically generate the network content. Under heavy system loads or any other unavailability or latency issues, the servers may be unable to provide the content. Accordingly, the requesting devices may receive error messages that the network content is unavailable. While the servers are experiencing issues, additional requests may continue to be received by the servers that can further delay recovery from the failure situations.